Skip to content
Back to skills

Pipeline

ASecurity

Full development pipeline: create GitHub issue, plan, implement with agent team, verify (build+test), commit, push, close issue. Use when you need to implement a feature or fix end-to-end.

  • 213 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added August 31, 2026
ai-agentsbashgit

Security analysis

A100/100

Scanned August 31, 2026

npx -y skills add spytensor/openmozi --skill pipeline --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Pipeline?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Pipeline
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/spytensor-pipeline/badge)](https://www.skillsdirectory.com/skills/spytensor-pipeline)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: pipeline
description: "Full development pipeline: create GitHub issue, plan, implement with agent team, verify (build+test), commit, push, close issue. Use when you need to implement a feature or fix end-to-end."
disable-model-invocation: true
---

# Development Pipeline

Automate the full development lifecycle for a feature or fix described by `$ARGUMENTS`.

## Workflow

Execute these steps in order. Stop and report if any step fails.

### Step 1: Create GitHub Issue

```bash
gh issue create --title "<concise title from $ARGUMENTS>" --body "<detailed description>"
```

Parse the returned URL to extract the issue number. Print it.

### Step 2: Plan

- Research the codebase to understand what files need changing
- Design the implementation approach
- Identify all files to modify/create
- Write the plan to a file at `.claude/plans/<issue-number>.md`

### Step 3: Build Agent Team & Implement

- Use `TeamCreate` to create a team named after the issue
- Break the plan into independent tasks using `TaskCreate`
- Spawn agents in parallel using the `Task` tool with `team_name` parameter:
  - Each agent handles one task (one file or one concern)
  - Use `subagent_type: "general-purpose"` with `mode: "bypassPermissions"`
  - Run agents in background when possible
- Wait for all agents to complete by checking `TaskList`

### Step 4: Review

- Read the modified files to verify changes are correct
- Check for obvious issues, missing imports, type errors

### Step 5: Verify

Run build and tests:
```bash
pnpm build
pnpm test
```

If build or tests fail:
1. Read the error output
2. Fix the issues directly
3. Re-run verification
4. Repeat until passing

### Step 6: Commit & Push

```bash
git add <changed files>
git commit -m "feat: <description> (#<issue-number>)

Co-authored-by: Mozi <MoziAI-co@users.noreply.github.com>"
git push origin main
```

### Step 7: Close Issue

```bash
gh issue close <issue-number> --comment "Implemented in <commit-hash>."
```

### Step 8: Cleanup

- Shut down all team agents via `SendMessage` with `type: "shutdown_request"`
- Delete the team via `TeamDelete`
- Report final summary: issue number, commit hash, files changed, test results

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.